Re: Complex CONTEXT index

>>Don't you need to translate the BLOB content into indexable text before you
>>index it? A simple transliteration of hex values is no help; you need
>>something that would convert the enclosed encoded Word or PDF into real
>>words.
>>
>>
>
>This is exactly what Oracle's Ultrasearch does. I've used it in the past
>(10gR1) to power an Intranet search site that crawled Intranet sites as well
>as file shares. It indexes very well, grabbing text from every popular
>format including binary MS documents, PDFs, drawings, as well as the headers
>in images and video files. But it wasn't the most stable. I needed to
>bounce it regularly, at least monthly, IIRC.
>
>Perhaps it's better with 10gR2 or 11g, if it's still available.
>
>HTH! GL!
>
>Rich
>
>
>--
>http://www.freelists.org/webpage/oracle-l
>
>
>
>

Dear All,
maybe not exactly what the OP is looking for, but why concatenate the different columns manually?
You can us the MULTI_COLUMN_DATASTORE to create a virtual document for each record.
I used it with VARCHAR2 and BLOB's as well as CLOB's. Please see link below for details:
http://download.oracle.com/docs/cd/B19306_01/text.102/b14218/cdatadic.htm#sthref327

Have a nice day
Manuela

--
http://www.freelists.org/webpage/oracle-l